ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

1 Variability of the Apparent Respiratory Quotient of a forest Soils and Tree 2 Stems 3 Boaz Hilman1,2, Tal Weiner1, Tom Haran1, and Alon Angert1

4 1 The Institute of Earth Sciences, The Hebrew University of Jerusalem, Givat-Ram, Jerusalem 5 91904, Israel 6 2 Department of Biogeochemical Processes, Max-Planck Institute for Biogeochemistry, Jena, 7 07745, Germany 8 Corresponding author: Boaz Hilman ([email protected])

9 Key Points:

10 • The ratio of CO2/O2 fluxes in (ARQ) depends on the substrate stoichiometry 11 and on additional processes reacting with the gases 12 • Bulk-soil ARQ value is governed by stoichiometry in high respiration rates and by 13 processes like abiotic Fe2+ oxidation in low rates 14 • ARQ can be used to partition soil respiration sources as shown by distinct bulk-soil and 15 roots ARQ values that confine total soil value 16

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

17 Abstract

18 The CO2/O2 fluxes ratio (ARQ) measured in soils and plants contains valuable information about 19 the respiratory-substrate stoichiometry and biotic and abiotic non-respiratory processes reacting 20 with the gases. For meaningful use in biogeochemical studies it is necessary to resolve the 21 substrate and processes effects. In addition, unique ARQ signatures can be used to 22 contributions to soil respiration. We investigated these uses by measurements in soil pore space 23 air (ARQsa), and in headspace air from incubations of bulk-soil (ARQbs) and tree stem-tissues 24 (ARQts for fresh tissues; ARQts24 after 24-h storage) in 10 measurement campaigns over 15 25 months in a Mediterranean oak forest. Mean (range) values were: ARQsa = 0.76 (0.60-0.92), 26 ARQbs = 0.75 (0.53-0.90), ARQts = 0.39 (0.19-0.70), and ARQts24 = 0.68 (0.42-1.08). Both 27 ARQts and ARQts24 were below 1.0, the value expected for respiration in plants. 28 Involvement of non-respiratory processes like non-phototrophic CO2 re-fixation and wound- 29 response O2 uptake (for ARQts) can explain the results. The mean ARQbs (0.75) probably 30 represents the stoichiometry of the respiratory substrate, which is lower than expected using bulk 31 soil organic matter (SOM) stoichiometry (~0.95), suggesting a labile, less oxidized, SOM pool 2+ 32 contributes more to respiration fluxes. Abiotic O2 uptake by Fe was demonstrated to reduce 33 ARQbs to 0.37, at the most, but estimated to have small effect under typical respiration rates. 34 ARQsa was usually higher than ARQbs and lower than root ARQ (which, when measured, ranged 35 from 0.73-0.96), demonstrating the potential of ARQ to partition the autotrophic and 36 heterotrophic sources of soil respiration. 37 38 Plain Language Summary 39 dioxide is produced by the processes of both metabolic respiration by plants and 40 microbial respiration and decomposition in soils. These are among the most important processes 41 in terrestrial ecosystems, both oxidizing organic compounds using O2 and emitting the resulting 42 CO2 to the atmosphere. However, our understanding of these processes, which are often 43 collectively referred to as “respiration” is still incomplete. Here we investigated the use of the 44 measured ratio CO2 released to O2 consumed, termed the apparent respiration quotient (ARQ), to 45 investigate respiration in soils and tree stems. ARQ measurements are rarely made, but can 46 provide valuable information about the chemistry of the respiratory substrates, and about 47 additional processes that involve CO2 and O2 cycling in terrestrial ecosystems. The expected 48 substrates () in tree stems and soils yield ARQ ≈1; however, we measured 49 considerably lower values. The low ARQ values in the soil can be explained if microbes 50 decompose compounds with low amounts of , which is surprising. No substrates can 51 produce ARQ values as low as those we measured in stem core incubations, indicating another 52 processes at work. 53 54 55 56 57

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

58 1 Introduction 59 New measurement methods can improve the understating and prediction of ecosystem processes. 60 An emerging tool in biogeochemical studies is the coupled measurement of CO2 and O2 fluxes. 61 In respiration the ratio CO2 production/O2 consumption is termed the ‘respiratory quotient’ (RQ). 62 The inverse term ‘oxidative ratio’ (OR, 1/RQ) is also used, especially when describing 63 ecosystem processes that include with O2 production (where the fluxes direction 64 is opposite to respiration). Both ratios depend primarily on the stoichiometry, of the respiratory 65 substrate or the net synthesized biomass. The stoichiometry of organic molecule determines the 66 mean oxidation state (Cox) of the C atoms in the molecule [LaRowe and Van Cappellen, 2011; 67 Masiello et al., 2008]. The more oxidized (higher Cox) the molecule, the fewer moles of O2 are 68 consumed per of CO2 released during complete oxidation and the RQ is higher (Table 1). 69 For this reason RQ is often used to infer which respiratory substrate is being used by plants and 70 animals. However, in soils and isolated plant organs additional biotic and abiotic non-respiratory 71 processes can influence CO2 and/or O2 and lead to a ratio of CO2/O2 that differs from the change 72 the substrate’s RQ value. For this reason, we refer to the ratio CO2 efflux/O2 uptake measured in 73 tree stems and soils as the ‘apparent RQ’ (ARQ) [Angert and Sherer, 2011; Angert et al., 2015]. 74 The main processes with the potential to affect ARQ are presented in Figure 1 and detailed in 75 section 1.1. In addition, in a recent study we suggested that ARQ measurements could be used to 76 separate autotrophic and heterotrophic respiration sources in soils due to different ARQ 77 signatures [Hicks Pries et al., 2019]. Thus, ARQ has the potential to become a useful tracer, 78 similar to δ13C in the ability to identify respiratory substrates, processes, and to separate 79 respiration sources [M E Gallagher et al., 2017]. 80 For meaningful use of ARQ as a biogeochemical tracer, it is crucial to identify and 81 separate the signals related to substrate stoichiometry and to non-respiratory processes. In 82 addition, the sources of ARQ variability in soils and tree stems are still unclear. Addressing these 83 questions has become more feasible with the development of user friendly O2 sensors based on 84 fuel-cell and optical quenching technologies. In the study we explore ARQ variability 85 and its sources in soils and tree stems of a Mediterranean oak forest, and discuss next research 86 steps. 87 88 89 90 91 92 93 94 95 96 97

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

98 Table 1 – Selected Compounds and their Type, Formula, Mean Oxidation State of the C (Cox), 99 RQ, OR, and the Gibbs Energy. a b c Compound Type Formula Cox RQ (ARQ) OR Gibbs energy (kj mol C-1)d

Oxalic acid Organic acid C2H2O4 3 4 0.25 -25.2

Glucose Carbohydrate C6H12O6 0 1 1 60.3

Cellulose Structural (C6H10O5)n 0 1 1 60.3 polysaccharide

Aspartic Amino acid C4H7NO4 1 1.33 (NH ) 0.75 (NH ) 31.8 3 3 acid 0.80 (HNO ) 1.25 (HNO ) 3 3

e Lignin Structural C100H124O43 -0.38 0.91 1.10 71.13 polymer

Oleic acid C18H34O2 -1.66 0.71 1.42 107.8 ()

Methane Inorganic gas CH4 -4 0.5 2 174.3 100 a The mean oxidation state of C in the molecule, known also as NOSC (the average nominal 101 oxidation state of carbon) calculated by C = where the coefficients are according to z 102 the molecule formula CaHbNcOd while z corresponds to the net charge [LaRowe and Van 103 Cappellen, 2011; Masiello et al., 2008]. This calculation assumes all non C species have fixed b 104 oxidation state during formation or break down. Cox is related to RQ by RQ = . The 105 respiratory quotient (RQ), the ratio CO2 efflux / O2 uptake in full aerobic oxidation of the 106 compound. We use the term ARQ (apparent RQ) since in soils and tree stems measurements 107 additional processes can cause deviation from the theoretic RQ. When molecule contains N the 108 RQ and OR values depend on the N oxidation state in the respective end and source molecules c 109 (e.g. NH3 (-3) or HNO3 (+5)). The oxidative ratio (OR = 1/RQ). Often viewed as the ratio O2 d 110 produced/ CO2 assimilated required for the biosynthesis of the compound. The Gibbs energy 111 for oxidation half reactions at 25oC and 1 bar calculated using LaRowe and Van Cappellen’s e 112 (2011) Eq. 14: G = 60.3 − 28.5 ∗ C. Lignin formula is according to Baldock et al. 2004.

113 1.1 Observed ARQ in soil respiration components and tree stems 114 The ARQ of soil heterotrophic respiration is usually approximated by root-free bulk-soil 115 incubations (ARQbs). According to meta-analysis of Cox in soil organic matter (SOM) [Worrall et 116 al., 2013] ARQbs values are expected to range between 0.77 and 1.11 with median of 0.95. 117 However, values of 0.27-0.94 measured previously from a variety of natural ecosystems and 118 agricultural lands are mostly below these expected values [Angert et al., 2015; Aon et al., 2001a; 119 b; Dilly, 2001; 2003; Dilly and Zyakun, 2008; Severinghaus, 1995]. Reasons for measured ARQ 120 in incubations to differ from the values expected from Cox of bulk soils include (Fig. 1): (1) 121 processes involving electron transfer that affect either O2 uptake or CO2 production without 122 affecting both gases; (2) interactions involving the greater of CO2 in water and 123 inorganic C cycling that do not affect O2 simultaneously; (3) other processes that affect CO2

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

124 more than O2 or vise versa such enzymatic uptake of CO2 by ‘dark fixation’ that affect CO2 more 125 than O2. 2+ 2+ 126 Enhanced O2 uptake derived from abiotic oxidation of reduced species like Fe and Mn 127 increases the denominator of the ARQ ratio and thus decreases its value. The opposite effect on 128 ARQ is expected during anoxic conditions when oxidized Fe3+ and Mn3+ for example are used as 129 an alternative electron acceptors. In that case, CO2 is emitted without any O2 uptake and the 130 numerator of the ARQ ratio increases. Anoxic conditions may exist within soil aggregates even 131 in aerated soils [Druschel et al., 2008; Hall and Silver, 2013; Sexstone et al., 1985], but become 132 more important after soil wetting as in water is slower by orders of magnitude than 133 diffusion in air, and when respiration rates are high and O2 replenishment in microsites cannot 134 meet respiratory needs. - 2- 135 reacts with water to form weak acids (HCO3 and CO3 ) and is ~30 times 136 more soluble than O2. Storage of respired CO2 as dissolved inorganic carbon (DIC) in soil water 137 therefore can also lower the measured ARQbs, especially in soils that may contain calcium - 2- 138 carbonate. At higher soil pH the of HCO3 and CO3 increase, and therefore also 139 the capacity of the soil water to hold DIC. However, if the water containing the DIC does not 140 leach, the CO2 is expected to degas back to the soil pore space when the soil is dried. In 141 calcareous soils mainly in arid and semi-arid regions large ARQsa deviations are expected due to 142 precipitation and dissolution of carbonates [Angert et al., 2015; Benavente et al., 2010; Cuezva et 143 al., 2011; Emmerich, 2003; T M Gallagher and Breecker, 2020; Ma et al., 2013].

144 Other processes involving CO2 and O2 include root uptake, SOM oxidation processes, 145 and dark fixation. DIC (or CO2) uptake by roots is probably not substantial due to anatomical 146 features [Ubierna et al., 2009]. ‘Oxidative depolymerization’ is the addition of oxygen to 147 macromolecules that forms smaller and soluble molecules, which are more readily digested by 148 microbes [Kleber et al., 2015]. If the oxygen-rich molecules are not respired and rather bind to 149 mineral surfaces, there is a net O2 uptake and ARQ is lowered. Dark fixation of CO2 by the 150 microbial community is another process that can lower ARQbs [Akinyede et al., 2020; Miltner et 151 al., 2005]. 152 In plants, the primary respiration substrates are carbohydrates with ARQ = 1.0 [Hoch et 153 al., 2003; Plaxton and Podestá, 2007]. This value is expected for roots, but measurements of 154 instantaneous ARQ from roots (ARQroot) often deviate from 1.0, ranging from 0.79 -1.4 155 [Hawkins et al., 1999; Rachmilevitch et al., 2006; Shane et al., 2004] (Fig. 1). These studies 156 measured ARQ from roots (from intact plants or excised) submerged in nutrient . 157 ARQroot values greater than 1.0 were explained by nitrate assimilation that consumes electrons 158 otherwise delivered to O2 [Bloom et al., 1989; Lambers et al., 2008; Rachmilevitch et al., 2006], 159 or by and lipid synthesis in the roots themselves or in the associated mycorrhiza, since 160 the conversion of carbohydrates to more reduced compounds result in ARQ >1.0 [De Vries et al., 161 1974; Hawkins et al., 1999; Shane et al., 2004]. Another process suggested to occur in roots is 162 dissolution of a fraction of the root-respired CO2 in the xylem water and its transport to above 163 ground tissues in the transpiration stream [Aubrey and Teskey, 2009; Grossiord et al., 2012]. 164 This CO2 removal is expected to reduce ARQroot during the daytime when transpiration is active. 165 The ARQ associated with respiration in the rhizosphere also depends on the composition of the 166 root exudates, which vary greatly in Cox [Bais et al., 2006]; ARQ will be above 1.0 when 167 exudates are dominated by organic acids and below 1.0 when dominated by amino acids (Table 168 1, Fig. 1).

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

169 Respiration in tree stems is also based on carbohydrates with expected ARQ (ARQts) of 170 1.0. However, the mean ARQts measured from stem chambers in tropical, temperate, and 171 Mediterranean trees was 0.59 and values of 1 were rarely observed [Angert et al., 2012; Hilman 172 et al., 2019] (Fig. 1). Dissolution and transport of respired CO2 via the xylem water stream is 173 thought to influence the CO2 efflux measured from tree stems [Teskey et al., 2008]. However, 174 CO2 transport was found to have only a minor role in explaining low ARQts in the investigated 175 trees [Hilman et al., 2019]. One indication for little effect of CO2 transport in the xylem water 176 was that ARQts values measured in stem chambers were similar to those measured during 177 incubation of stem tissues that were clearly detached from the xylem system and unaffected by 178 transport. An alternative hypothesis for lower than expected ARQts values is non-phototrophic 179 CO2 fixation by the enzyme phosphoenolpyruvate carboxylase (PEPC) [Hilman et al., 2019], 180 which was found to be highly abundant in young tree stems [Berveiller and Damesin, 2007; 181 Berveiller et al., 2007]. Photosynthetic fixation of CO2 under tree bark would produce O2 and 182 would not produce ARQ that deviates from 1.0.

183 184 Figure 1 Observed ARQ in different ecosystem components with plausible explanations for 185 deviations from the expected value. The expected ARQ is according to the putative-substrate 186 stoichiometry (see Table 1): Carbohydrates for respiration in roots and stems; rhizodeposition 187 composition for rhizosphere respiration, with values indicative of the end-members amino and 188 organic acids; and soil organic matter composition for root-free soil respiration according to 189 Worrall et al. [2013]. The observed ARQ values are taken from the literature: Stem respiration 190 [Angert et al., 2012; Angert and Sherer, 2011; Hilman et al., 2019]; Root respiration [Hawkins et 191 al., 1999; Rachmilevitch et al., 2006; Shane et al., 2004]; Root-free soil respiration [Angert et al., 192 2015; Aon et al., 2001a; b; Dilly, 2001; 2003; Dilly and Zyakun, 2008; Severinghaus, 1995]; Soil 193 pore air, tube sampling [Angert et al., 2012; Angert et al., 2015; Hicks Pries et al., 2019; 194 Sanchez-Canete et al., 2018], soil chambers [Ishidoya et al., 2013; Seibt et al., 2004]. Included

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

195 also are the expected effects on ARQ and the photosynthetic ratio (O2 produced/ CO2 196 assimilated) if C is imported internally to the canopy.

197 1.2 ARQ as a tool to separate respiration sources in soil and ecosystem 198 Total soil respiration integrates heterotrophic respiration in root-free soil and the root-associated 199 rhizosphere with autotrophic respiration by roots. Measurements of ARQ of total soil respiration 200 (ARQsa) have been made in two ways. Most frequently, they have been estimated from the 201 difference in the ratio of CO2/O2 in soil air pore space (sampled by a tube) compared to the 202 CO2/O2 of overlying air, corrected for diffusivity differences [Angert et al., 2015]. Using this 203 method a wide range of values, between 0.23 and 1.14, were measured in a variety of forest soils 204 from different biomes [Angert et al., 2012; Angert et al., 2015; Hicks Pries et al., 2019; Sanchez- 205 Canete et al., 2018]. The second measurement monitored simultaneous changes in CO2 and O2 in 206 the headspace of a chamber covering the soil surface. These measurements tend to have higher 207 ARQsa values, ranging between 0.90-1.06 [Ishidoya et al., 2013; Seibt et al., 2004]. The large 208 observed ARQsa variability can be attributed to two factors: variability in the relative weight of 209 each of the soil respiration sources, and variability within each of the sources. For example, 210 Hicks Pries et al. [2019] found strong seasonality in ARQsa in western US forest conifer stand 211 with summer vs. winter values of 0.89 and 0.70, respectively. The seasonal variation, observed 13 212 also in soil air δ CO2, was attributed to changes in respiratory substrates that switched between 213 root-based respiration of more oxidized compounds during summer and bulk-soil respiration of 214 more reduced compounds during winter (Table 1, Fig. 1). Similarly, greater contributions of 215 autotrophic respiration and litter decomposition in shallower horizons that are not well-captured 216 by soil gas sampling tubes might also explain the high ARQsa measured from soil chambers. 217 Overall, different ARQ values for different components of soil (or ecosystem) respiration show 218 promise for its use to partition respiration sources or identify transport of CO2 between 219 components.

220 While measured values of ARQ from soils and stems are often < 1, total ecosystem CO2 221 and O2 fluxes are nearly balanced (ARQ ~ 1). Direct estimates of the oxidative ratio (OR = 222 1/ARQ) using gas measurements in and above canopies ranged between 0.94 and 1.10 over diel 223 and annual cycles [Battle et al., 2019; Ishidoya et al., 2013; Seibt et al., 2004; Stephens et al., 224 2007]. OR over longer time scales can be estimated by the ecosystem’s Cox [Keeling, 1988; 225 Masiello et al., 2008; Severinghaus, 1995]. According to a meta-analysis of total biomass and 226 SOM-Cox, OR is mostly between 1.00 – 1.13 in 16 global biomes [Worrall et al., 2013]. 227 Therefore, the low ARQ measured in soils and tree stems must be balanced, at least over longer 228 time scales, by higher ARQ measured elsewhere.

229 The fact that ARQts is almost always lower than the expected value of 1 from substrate 230 carbohydrates may indicate a process that retains respired CO2 within the stem [Hilman et al., 231 2019]. If the CO2 retention in the stem is due to PEPC re-fixation, biosynthesis of compounds 232 more oxidized than carbohydrates like organic acids are expected [Lambers et al., 2008]. The 233 catabolism of organic acids results in ARQ >1 (Table 1) that will balance the observed low 234 ARQts where CO2 is fixed. The high ARQ may appear in the rhizosphere, if organic acids are 235 exported via the phloem to the roots and transferred to the soil as root exudates [Hoffland et al., 236 1992; Shane et al., 2004]. Alternatively, the organic acids might be transported upwards in the 237 xylem water [Schill et al., 1996] and contribute to respiration in the canopy. For example, the

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

238 organic acid malate can contribute C to leaf photosynthesis [Hibberd and Quick, 2002], and alter 239 the photosynthetic ratio evaluated by leaf O2/CO2 fluxes.

240 Another process proposed for CO2 retention in the stem is CO2 dissolution in the xylem 241 water. In this case, CO2 degassing in higher parts of the tree will result in ARQ values that are 242 locally >1.0. In addition, dissolved CO2 may contribute to canopy photosynthesis [Bloemen et 243 al., 2013; Stringer and Kimmerer, 1993] in which case it will also affect the photosynthetic.

244 2 Materials and Methods 245 In this study, we focused on the seasonal variation of ARQ in soils and tree stems. Seasonal 246 measurements at ~1.5 months intervals of in situ ARQsa estimated from soil gas tubes and a 247 diffusion model, and ARQbs and ARQts using soil and tree stem incubations were conducted in a 248 Mediterranean oak system. After initially performing bulk-soil incubations (ARQbs) at room 249 , we realized that the different in the field may affect measured ARQbs. 250 We therefore performed an additional experiment to test the temperature effect on ARQbs and 251 used this to correct the ARQbs results. The evaluation of the control of physical variables (water 252 availability and temperature) on ARQ seasonal variability was performed using a backward 253 selection technique for multiple regressions. ARQsa was expected to be higher than ARQbs due to 254 the contribution of root respiration with higher ARQ than in the bulk soil. A direct test was done 255 by comparison of ARQsa, ARQbs, and ARQroot sampled at the same time. Additional experiments 2+ 2+ 256 were conducted to investigate the potential of Fe and Mn oxidation to reduce ARQbs. 257 Following the hypothesis that low ARQts is the result of CO2 re-fixation and production of 258 organic acids, we expected anti-correlation between ARQts and ARQsa if the produced organic 259 acids are secreted finally to the rhizosphere as exudates. 260

261 2.1 Study site 262 The study took place in Odem Forest, located 950 m a.s.l, 33o13’ N, 35o45’ E. The climate is 263 Mesic Mediterranean with a mean annual precipitation of 950 mm and summer and winter mean 264 temperatures of 21.3º C and 7.3ºC, respectively. The dominant tree species are the evergreen 265 Quercus calliprinos Webb (about 75% of the woody cover area) and the winter-deciduous 266 Quercus boissieri Reut. (15%) [Kaplan and Gutman, 1996]. Q. calliprinos is the dominant tree in 267 the Mediterranean scrub in Israel, while Q. boissieri grows mainly in altitudes above 500 m a.s.l 268 [Kaplan and Gutman, 1996]. The soil was formed on basaltic bedrock and is classified as Eutric 269 Lithosol in the FAO classification system and as Lithic Xerorthent in the USDA classification 270 system. The soil pH is 6.6 and the organic C content in the top 10 cm is 12% [Gross and Angert, 271 2017].

272 2.2 Experimental design

273 2.2.1 Seasonal measurements 274 Seasonal sampling took place in ten campaigns between February 2017 and May 2018. Soil air 275 was sampled from 1/2" (OD) stainless steel tubes closed at the bottom end, and perforated near 276 the bottom, that were hammered into the soil. The samples of soil air were collected from a depth 277 of 15 ± 4 cm in pre-evacuated ~3.6 mL glass flasks with Louwer™ O-ring high-vacuum valves. 278 Before sampling, the dead volume in the tubing and flask necks was purged with soil air by a

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

279 plastic syringe equipped with a two-way valve. A total of 120 samples were taken near each tree 280 species (2 replicates x 2 samples x 3 trees x 10 campaigns). Every tree was sampled only once 281 since sampling caused some disturbance to the soil and the stem (see below). 282 Surface soil from 0-10 cm depth was collected with a trowel and stored in a plastic bag. 283 A total of 30 samples were taken near each tree species by pooling from two places near each 284 tree (3 trees x 10 campaigns). Soil moisture was measured gravimetrically on ~3 g subsamples 285 (available only for the last 6 campaigns). For bulk soil incubation experiments, the soil was 286 sieved to 2 mm (except on January 2018 sampling when the soil was too wet and sticky to allow 287 sieving), and a subsample of 3 g was incubated overnight in 6 mL glass test tubes connected to 288 ~3.6 mL glass flasks by Ultra-Torr fittings (Swagelok, Solon, OH, USA). The gas in the 289 headspace had initial mean atmospheric values (20.95% O2, 0.04% CO2). Incubations were 290 conducted usually two days after soil collection at room temperature.

291 For estimating ARQts we performed stem tissue incubations. This method was shown to 292 give similar ARQ values as the stem-chamber method in three species including the oak Quercus 293 ilex [Hilman et al., 2019]. We decided to incubate only the phloem and cambium tissues since 294 they are the most metabolically active tissues in the stem [Bowman et al., 2005], and since 295 transport in the phloem is the pathway for C to flow from the stem to the roots. Cores were 296 extracted using a 1.0 cm diameter cork borer, at 20 cm and 130 cm above the soil surface. A total 297 of 60 samples were taken from each tree species (2 stem positions x 3 trees x 10 campaigns). We 298 removed from the cores the outer bark and sapwood sieves, and further cut the cores to fit into 299 the 3.6 mL glass flask neck. For the incubations, we plugged the neck with a rubber stopper to 300 create a gas-tight headspace with initial mean atmospheric values. The incubations started 301 immediately after harvesting and lasted 3-4 hours in the dark and at environmental temperatures. 302 in stem cores changes rapidly after harvesting; in a previous study ARQts increased 303 with time after harvesting from 0.4 to values closer to 1.0 while the O2 uptake rate was 304 maintained, indicating an increase in CO2 efflux over time [Hilman et al., 2019]. The increased 305 ARQts may provide evidence for gradual inhibition of PEPC activity that fixed respired CO2 306 early during the incubation, but became less efficient as its own products accumulated. To 307 observe potential temporal changes in ARQts, the tissues were re-incubated 24 hours after 308 harvesting (ARQts24) for the same duration at room temperature. In the 24 h storage time stem 309 tissues were wrapped with moist gauze cloth to avoid desiccation.

310 2.2.2 Comparison of roots, bulk soil, and soil air ARQ and temperature effect 311 To evaluate the potential for ARQ to partition the contributions of autotrophic and heterotrophic 312 sources to total soil respiration, and to test the sensitivity of ARQbs to incubation temperature, we 313 sampled additional trees. During January 2019 we measured ARQsa, ARQbs, and ARQroot near 314 three additional trees from each species. For ARQroot fine roots (< 2 mm), which showed the 315 highest specific respiration rates among the root diameters tested, were excavated [Chen et al., 316 2010; Desrochers et al., 2002; Pregitzer et al., 1998]. To test possible effects of root-surface 317 microbial communities on ARQroot, soil was washed thoroughly from one subsample of roots 318 before incubation (n = 3), while a second subsample (n = 3) was incubated with the surrounding 319 soil. Roots were incubated shortly after harvesting in the dark in a set-up of two 3.6 mL glass 320 flasks connected by Ultra-Torr fitting, and kept at ~7oC to represent field condition (field soil 321 temperature was 6-8oC). Since we expected low respiration rates incubations lasted 24 h. We 322 also included in the comparison coarse roots of Q. calliprinos (< 1 cm in diameter) collected in

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

323 March 2018. Bulk soil incubations were conducted at temperatures of 6, 22, and 30 oC and lasted o 324 68-90 h (2 samples x 3 trees). The Q10, the factor by which respiratory flux rises with a 10 C 325 increase, was calculated using the function Q10 from the R package respirometry [R Core Team, 326 2019]. We also present ARQbs values for soils sampled in March and May 2018, when soil 327 temperatures were 1°C and 22oC, respectively (n = 1).

328 2.2.3 Evaluation of the effect of abiotic O2 uptake on bulk soil ARQ 329 Two additional soil incubation experiments were undertaken to investigate the potential for 330 abiotic O2 uptake to affect ARQbs. In the first experiment we tested the response to temporary 331 anaerobic conditions with un-screened soils (for maintaining their structure). Mason jars (1 L) 332 with a small volume of soil (~150 ml, n = 3) and jars with large soil volume (~550 ml, n = 3) 333 were incubated for 13 days, to create low O2 concentrations. Headspace [O2] was measured by 334 the end of the incubation, and soils were sampled for [Fe2+] and [Mn2+]. The soils were then 335 ventilated for 1.5 hours, before an overnight incubation. Air and soil samples were measured 336 again at the end of this incubation. The soil moisture measured at the beginning of the 337 experiment was 31% by weight. The soil [Fe2+] was measured by the Ferrozine method [Liptzin 338 and Silver, 2009]. The soil samples were sieved to 2 mm, and extracted by 0.5 M HCl 339 immediately at the end of the incubation experiments. The soil [Mn2+] was measured by 340 assuming that HCl-extractable Mn, which was quantified by ICP (7500cx Agilent technologies, 341 Santa Clara, CA, USA), predominantly represents Mn2+ [Keiluweit et al., 2018].

342 In a second experiment we tracked ARQbs during a wetting-drying cycle, and measured 343 [Fe2+] and soil moisture during the soil drying. Ultra-Torr Tee fittings (Swagelok, Solon, OH, 344 USA) were used for the incubation, connecting a test-tube with soil, a test-tube with a drying- 345 agent (magnesium perchlorate), and a 3.6 ml flask equipped with Louwer™ O-ring high-vacuum 346 valve. We incubated 2-mm sieved soil (n = 1) and un-sieved soil (n = 1). After every incubation 347 the flask was closed and removed, the system was ventilated for 1 hour and then new flask was 348 attached. The first incubation was used to determine the basal ARQbs and respiration rate (O2 349 uptake). The soil was then dried for 17 days, wetted, and dried again for 26 days. Soil wetting 350 was roughly equivalent to a rainfall event of 20 mm. The destructive Fe2+ and soil moisture 351 measurements during the soil drying were done for the sieved soil, after re-wetting it to the same 352 degree. We report the relative respiration rate (RR) as the ratio between the O2 uptake in each 353 incubation to the basal rate.

354 2.4 Gas analysis

355 The [O2] and [CO2] of the air samples were measured in the laboratory by a closed system (The 356 “Hampadah” [Hilman et al., 2019]). The system is based on two analyzers: an infra-red gas 357 analyzer (IRGA) for CO2 measurement (LI 840A LI-COR; Lincoln, NE, USA) and a fuel-cell 358 based analyzer (FC-10; Sable Systems International, Las Vegas, NV, USA) for measuring O2, 359 and is fully automated.

360 For measuring [CO2] and [O2] from the Mason jars we equipped each lid with a septum. 361 Air from the headspace was sampled by plastic syringe with needle and injected into a flow- 362 through CO2 (K33 ICB 30% CO2 Sensor, CO2 Meter, Inc) and O2 (Fibox 3, PreSens-Precision 363 Sensing) sensors, connected by plastic tubing. The O2 sensor is a quenching based optical fiber 364 (optode) that reads the fluorescence from a sensing "spot". We placed the “spot” in a 3 mm clear 365 plastic aperture in an opaque lid of a custom-made 2-cm diameter flow-through cell, which made

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

366 from 4 mm thick aluminum base (to stabilized the temperature). From the outside of the aperture 367 a connector for the optical fiber that reads the "spot" fluorescence was fixed. The same air was 368 injected to pre-evacuated ~3.6 mL glass flasks for comparison with the "Hampadah" method.

369 2.3 ARQ and respiratory fluxes estimations

370 ARQsa, the CO2 efflux/O2 uptake in soil respiration, was calculated from the measured gases 371 concertation using the following equation [Angert et al., 2015]: ×([ ] [ ] ) ∆ 372 ARQ = = −0.76 × (1) ×([][]) ∆

373 Where Dco2, [CO2]s, and [CO2]a are respectively the effective diffusivity of gaseous CO2, 374 and the CO2 concentrations in the soil (measured value) and in the ambient air (assumed to be 375 0.04%). ΔCO2 is therefore the difference in [CO2] between the soil and the atmosphere. The 376 same definitions hold for O2, where the in the ambient air assumed to be 20.95%. 377 Since the O2 flux is opposite in direction to the CO2 flux we added negative sign for 378 convenience. The effective diffusivity D depends on the structure of the pore spaces and on the 379 diffusivity in air. It can be assumed that the structure is identical for CO2 and O2, therefore the 380 ratio Dco2/ Do2 depends only on the CO2/O2 diffusivity ratio in air, which is 0.76 [Massman, 381 1998]. 382 Eq. 1 is somewhat analogous to the Davidson [Davidson, 1995] equation that estimates 13 383 the respired δ CO2 in soil from soil-air discrete sample. The equation corrects for the faster 12 13 384 diffusion of CO2 (in our case O2) that enriches the soil air with the slower-diffusing CO2 (in 13 385 our case CO2), and for difference in δ C and [CO2] between the soil and the air above the soil. 386 The Davidson equation accounts for diffusion and diffusive mixing and thus yields the same 13 387 respired δ CO2 regardless soil depth [Bowling et al., 2015; Egan et al., 2019]. Eq. 1 is expected 388 to similarly account for gas-mixing effects with depth [Angert et al., 2015]. Indeed, ARQsa was 389 similar when measured at depths of 10, 30, and 60 cm (0.3, [Sanchez-Canete et al., 2018]) and at 390 30 and 90 cm (0.8, [Hicks Pries et al., 2019]). Moreover, as ΔCO2 and ΔO2 are calculated with 391 respect to atmospheric air, advective mixing of soil air with the above air is not expected to 392 change the measured ARQsa. Nonetheless, since such mixing will violate the diffusional steady- 393 state between different soil depths we avoided sampling in days with high wind speeds (> 4 m s- 394 1).

395 ARQbs, ARQts, and ARQts24 were calculated by the ratio between [CO2] and [O2] net 396 percent changes in the incubation headspace. ARQbs values were further corrected for CO2 397 dissolution since the large volume of water in the bulk soil samples collected during winter and 398 the fairly high pH value for non-calcareous soil (6.6) are expected to cause to some of the - 399 respired CO2 to convert into DIC. DIC (mmol/L), the sum of dissolved CO2, H2CO3, HCO3 , and 2- 400 CO3 , was calculated using Eq. 2 [Stumm and Morgan, 1996]: × 401 DIC = P ×k ×1+ + (2) ()

402 where PCO2 is the partial of gaseous CO2 (= [CO2]/100), kh is Henry’s constant 403 that states the amount of dissolved CO2 and H2CO3 in the water, and k1 and k2 are the first and 404 second acidity constants, respectively. The constants are temperature dependent and calculated 405 according to the incubation temperature [Harned and Davis, 1943; Harned and Scholes Jr, 406 1941]. To estimate the addition of DIC during incubation we assumed the initial DIC was in

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

407 equilibrium with atmospheric [CO2] of 0.04%, and the final DIC in equilibrium with the final 408 [CO2]. The net change in the calculated DIC (ΔDIC) was converted to gaseous CO2 equivalent 409 and added to the measured [CO2]measured, yielding the corrected [CO2]c: ∆ ××. 410 [CO] = [CO] + 100 × (3)

411 where W is the absolute amount of water in the sample (L) and VHS is the headspace 412 volume (mL). When soil moisture data were unavailable, we estimated its value from the relation 413 between the available soil moisture data and rainfall in the last 3 weeks. The term 22.4 converts 414 the units of the dissolved CO2 (mmol) to mL gas.

415 The O2 uptake and CO2 efflux from the bulk-soil incubations were calculated using the 416 equation (nanomole gas g.DW-1 min-1):

∆×× 417 Flux = (4) ×××.×

418 where ΔGas is net percent change in the gas concentration ([O2] or [CO2]) during the 419 incubation, BP is the local barometric pressure (hPa), t is the temperature (k), M is the soil dry -3 -1 420 weight (g), It is the incubation time (min), and 8.314×10 is the ideal gas constant (mL hPa k 421 nanomol-1). Soil samples were oven-dried (105oC, 24 h) for dry .

422 2.5 Statistical analysis 423 All comparisons were conducted using one-way analysis of variance (ANOVA) and a t-test, after 424 assuring homogeneity of variances using Bartlett’s test. For unequal variances we used a 425 Welch’s test and nonparametric comparisons with Wilcoxon method. Significant differences 426 were determined at P < 0.05. The relations of the seasonal ARQ and bulk-soil O2 uptake values 427 with physical parameters were tested using backward selection technique for multiple 428 regressions, including estimates of the interactions between each two factors. The tested 429 parameters were: soil moisture (available for the last 6 out of 10 campaigns), the number of days 430 passed since the last rain event, the accumulated rain in the 3 weeks prior to sampling, and soil 431 temperature at 10 cm depth. Rainfall and temperature were measured in the nearby metrological 432 station of El rom (data courtesy of Meteo-Tech Ltd. Meteorological Services). We used linear 433 regressions not only to evaluate the relationship of dependent and independent variables, but also 434 to describe correlation between ARQts and ARQsa. All statistical analysis was done using JMP 435 (JMP®, JMP Pro 13, SAS Institute Inc., Cary, NC, USA).

436 3 Results

437 3.1 Seasonal measurements 438 The seasonal ARQ measurements are presented in Figure 2. The overall mean ± SE values 439 (range of means per species per date) of ARQsa, ARQbs (without correction for CO2 dissolution 440 in water), ARQts, and ARQts24 were respectively 0.76 ± 0.02 (0.60-0.92), 0.65 ± 0.02 (0.47-0.80), 441 0.39 ± 0.03 (0.19-0.70), and 0.68 ± 0.04 (0.42-1.08). The dissolution correction for ARQbs values 442 increased the mean value and range to 0.72 (0.51-0.88). The correction had some sensitivity to 443 the parameters pH, W (soil water content), and VHS (incubation headspace) (Eq. 2-3). The 444 decrease of VHS in 10%, and the increase of W in 10% and the pH from 6.6 to 6.8 resulted in an 445 increase of the overall mean to 0.76. Varying the parameters in the same magnitude but in

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

446 opposite direction yielded overall mean of 0.69. The weighted mean of the corrected ARQbs 447 (using O2 uptake rates for weighting) further increased the mean value to 0.75 (0.53-0.90). From 448 this point on in the paper ARQbs values will refer to values corrected for CO2 dissolution. 449 Significant difference between species were observed only in ARQts (P = 0.0015, Welch test, 450 both stem heights pooled together) with mean values of 0.46 ± 0.03 and 0.32 ± 0.02 for Q. 451 boissieri and Q. calliprinos, respectively. Marginal significance (P = 0.06, t test) was observed 452 for the weighted ARQbs where Q. calliprinos was higher than Q. boissieri (0.77 vs. 0.72).

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

453

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

454 Figure 2. Seasonal measurements of (a) ARQsa ± SE, the ratio of CO2 efflux/O2 uptake measured 455 for soil air in depth of 15 ± 4 cm (n = 3); (b) ARQ measured from bulk soil incubation (ARQbs) 456 where full markers with soil line indicate the measured values and empty markers with dashed 457 lines indicate the dissolution corrected values (n = 3); (c) the mean ARQ measured for incubated 458 stem tissues (ARQts) extracted from at 20 and 130 cm above the ground (ARQts, full markers 459 indicate incubations conducted directly after the core was harvested; empty markers with dashed 460 lines indicate incubations conducted 24 h after harvest) sampled 20 and 130 cm above ground, 461 (d) the O2 uptake rate of the incubated bulk soils, (e) daily precipitation (black bars) and soil 462 temperature (blue line) measured by adjacent meteorological station and the soil moisture in the 463 site. Shaded periods indicate winter dormancy of the deciduous Q. boissieri. Soil sampling was 464 conducted underneath the trees. Error bars represents standard errors.

465 3.2 Bulk soil measurements

466 The seasonal variability of ARQbs was not explained by the tested physical parameters in the 467 backward selection technique. When temperature was tested individually it had positive effect on o o 468 ARQbs between 1-6 and 22 C, while between 22 and 30 C ARQbs values were rather stable (Fig. 469 3a). At the highest temperatures we observed species effect with higher values in the Q. 470 calliprinos (Fig.3a). The Q. calliprinos also had higher CO2 and O2 fluxes, greater sensitivity to 471 temperature (higher Q10 values, Fig. 3b,c), and greater soil moisture than the Q. boissieri (Fig. 472 3d). For correction of the seasonal incubations conducted at room temperature to field soil 2 473 temperature, we used a logarithmic fit (R = 0.78) equation for the relation between Tincubation (the o 474 temperature in which the incubation took place (C )) and ARQbs for both species:

475 ARQ =0.13×ln(T) +0.36 (5) 476

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

477 478 Figure 3. Results (mean ± SE) from bulk soil incubations at different temperatures. Filled 479 symbols represent soils collected on January 2019 underneath three trees from each species with 480 two soil samples per tree (n=6). Empty symbols represent soils collected on March and May 481 2018. (a) ARQbs (ratio of CO2 efflux/O2 uptake) with logarithmic fit, (b) the CO2 efflux rates 482 (after CO2 dissolution correction) and the calculated temperature coefficient Q10, (c) the O2 483 uptake rates and the calculated Q10 values, and (d) the gravimetric moisture of the soils. 484 Asterisks indicate significance difference between species (** - P < 0.01, *** - P < 0.001) in t

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

o 485 test, expect for the CO2 and O2 fluxes at 22 C where Welch test was used. For soil moisture 486 comparison was made for all temperatures together.

487 A trend of higher ARQbs values with higher bulk-soil O2 uptake rates was observed for 488 both the seasonal measurements and the temperature-controlled incubations (Fig. 4). In addition, 489 variability declined with increasing O2 uptake rates. The bulk-soil O2 uptake rate showed a 490 strong seasonal cycle, with maximal uptake rates during spring (March-May) and minimal rates 491 during the end of the summer (August-October) (Fig. 2d). The uptake rates of the two species did 492 not differ significantly (P = 0.766, t test). A reciprocal effect on bulk-soil O2 uptake rate -1 493 (nanomole O2 g.DW min ) was found between soil moisture (M) and soil temperature (Tsoil). 494 The effect is described by the following equation:

495 O uptake = 17.50 M + 0.09 T + (T−14.5) × (M−0.24) × 0.59 − 2.79 (6) 496 The actual respiration rate versus the equation predicted respiration gives R2 = 0.94 (P < 497 0.0001). A significant linear relation was found also between bulk-soil O2 uptake and the number 498 of days elapsed since the last rain event (R2 = 0.4, P = 0.005). Adding this effect to the prediction 499 formula does not improve R2 which remains 0.94 (no reciprocal effect was found in relation to 500 this parameter). A correlation coefficient R2 of 0.75 (P < 0.001) was calculated while assuming 501 M is the only driving factor on bulk-soil O2 uptake. 502

503

504 Figure 4. Scatter plot of ARQ (the ratio of CO2 efflux/O2 uptake) measured from bulk soil 505 incubations (ARQbs) and the O2 uptake rate of the incubated bulk soils. Results are grouped by 506 the month of sampling for the seasonal measurements (colored circles) and temperature

2+ 507 3.2.1 Relation between ARQbs and Fe

508 After the first 13-days of incubation the average [O2] ± SD of the incubation jars with the large 509 and small soil volumes were 0.90 ± 0.44% and 7.25 ± 0.07%, respectively (n = 3). In agreement, 510 [Fe2+] in the jars with large soil volume was higher than measured for the small soil volume jars 511 (0.89 ± 0.24 vs. 0.05 ± 0.01 mg g-1 soil, respectively). In the subsequent incubation, performed 512 after ventilation of 1.5 hours aimed to increase the [O2] in the jars, a sharp decrease in [O2] was 513 observed in the large soil volume jars from an ambient value of 20.95% to value of 4.77 ±

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

514 0.21%, while ARQ was 0.37 ± 0.01. The [Fe2+] dropped from 0.89 ± 0.24 to 0.21±0.04 mg g-1 515 soil. The [Mn2+] was 1.27 mg g-1 soil, and did not significantly change during this incubation. In 516 the small soil volume jars the [O2] decreased from 20.95% to 19.80 ± 0.26%, ARQ was 517 0.74 ± 0.02, and [Fe2+] did not change from the initial value of 0.05 mg g-1 soil. Taking into 518 account the different soil volumes, the rate of O2 uptake was 2.9-fold faster in the large soil 519 volume jars than in the small soil volume jars. The [CO2] and [O2] determined by the sensors in 520 this experiment were highly consistent with the Hampadah measurement (R2 of 0.997 and 0.975 521 in linear regression with slopes of 1.01 and 1.01, respectively). 2+ 522 The soil wetting-drying experiment induced variations in ARQbs, RR, and [Fe ] (Fig. 5). 523 RR peaked in the day of soil wetting and then gradually decreased. Following the soil wetting 524 ARQbs increased during 11 days from 0.63-0.69 to 0.79-0.80 and then decreased during 15 days 525 to 0.46. [Fe2+] values of ~0.14 mg g-1 were measured during the first 13 days after soil wetting, at 526 soil moisture values of 14.4%-7.7%. After the 13th day [Fe2+] decreased to 0.10 mg g-1 in the 34th 527 day and to 0.05 mg g-1 in the 46th day.

528 529 Figure 5. Results from soil drying-rewetting experiment. The day of the rewetting is day 0. (a) 530 ARQbs (ratio of CO2 efflux/O2 uptake) in solid lines and relative respiration rate (RR) in dashed 531 lines for un-sieved and sieved (2 mm) soils. Each data point represents one measurement without 532 replicates. (b) The concentration of Fe2+ (mg g-1) and the gravimetric moisture of the sieved soil. 533 Following the experiment presented in panel a, the same sieved soil was wetted to the same 534 moisture. Each data point represents mean of duplicate sub-samples taken from the drying soil. 535 Error bars are the standard deviations.

536 3.3 Soil air measurements

537 Concentrations of CO2 and O2 in the soils in single tube samplings ranged from 0.17 - 2.25% and 538 20.79 - 18.14%, respectively. The lowest O2 concentrations were measured during January 2018

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

539 after 163 mm of precipitation over the previous 3 weeks. The seasonal variability of ARQsa was 540 explained by the water related parameters M, the number of days passed since the last rain event 541 (D), and accumulated rain in the 3 weeks prior to sampling (R) in the backward selection 542 technique. A reciprocal effect was found between the last two factors. The statistical model is 543 defined by the equation: 544 ARQ = 0.47M + 0.02D + 0.004R + (D−18) × (R−58.5) ×3×10 +0.24 (7) 545 With P = 0.0002 on F test and the correlation between the actual and predicted soil ARQ 2 546 gives R of 0.94. The effect of Tsoil is small testing all-year time scale and its addition to the 547 prediction formula has a minor contribution to the coefficient of determination. However, when 548 omitting from the analysis data collected during late winter and spring and including only data 2 549 from May 2017 – Jan 2018, ARQsa is found to be strongly dependent in temperature (R = 0.92, 550 P < 0.0001). The relation is given by the linear equation: ARQsa = 0.01×Tsoil + 0.6.

551 3.4 Comparison of ARQ in bulk soil, roots, and soil air

552 The mean ARQroot per species per date ranged between 0.73-0.96 (Fig 6a). The root washing 553 treatment in roots collected in January 2019 did not affect ARQroot (P = 0.99, t test) therefore 554 results were pooled together. ARQsa values were always lower than ARQroot and higher than 555 ARQbs (Fig. 6a). Assuming root and bulk-soil respiration are the only end members affecting the 556 soil pore space air, their relative contributions could be estimated using a simple mixing model, 557 where ARQsa = X × ARQroot + (1 - X) × ARQbs (Fig. 6a).

558 The seasonal ARQbs measurements were conducted in room temperature and not in the 559 actual field soil temperature. Thus, for comparison of ARQbs and ARQsa (measured at field 560 conditions) ARQbs has to be corrected to field temperature. We first averaged values of both 561 species at each measurement date. Then the intercept term b from Eq. 5 was modified:

562 b= 0.36−(ARQ − ARQ_) (8)

563 Where 0.36 is the calculated intercept as appears in Eq. 5, ARQt=room is the expected 564 ARQ according to Eq. 5 and the room temperature (varied slightly between measurements), and 565 ARQbs_measured is the measured ARQ in the bulk-soil incubation, corrected to CO2 dissolution. 566 The bulk soil ARQ values reported in Figure 7b were calculated with the equation:

567 𝐴𝑅𝑄 = 0.13 ×𝑙𝑛(𝑇) +𝑏 (9)

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

568 569 Figure 6. (a) A comparison of ARQ (ratio of CO2 efflux/O2 uptake) values (mean ± SE) 570 measured from root incubations (n = 2, 6, 6), soil air (n = 3), and bulk-soil incubations (n = 3). 571 The x axis indicates the date of sampling and the tree species. The relative contributions (%) of 572 roots and bulk-soil respiration to the total soil respiration are indicated in the bars. The 573 contributions were calculated using the equation ARQsa = X × ARQroot + (1 - X) × ARQbs. (b) 574 The seasonal course of ARQ means (± SE, n = 6) of both tree species, where the bulk soil values 575 are temperature corrected according to Eq. 9. Shaded periods indicate winter dormancy of the 576 deciduous Q. boissieri.

577

578 3.5 Tree stem measurements

579 The strongest effect on ARQ measured in tree stems was time; ARQts values measured in 580 incubations started immediately after tissues extraction were much lower than ARQts24 measured 581 in incubation started 24 h after the extraction (Fig. 2c). The overall ARQts and ARQts24 means 582 (±SE) for the Q. boissieri were 0.46 ± 0.03 and 0.68 ± 0.03, respectively (P < 0.0001, Welch 583 test), and for the Q. calliprinos 0.32 ± 0.02 and 0.67 ± 0.03, respectively (P < 0.0001, Welch 584 test). The sampling height of the stem cores (20 and 130 cm) had no significant effect when 585 analyzed separately for ARQts and ARQts24 in both species (P = 0.14 at the minimum). 586 In general, the deciduous Q. boissieri exhibited greater seasonal variability than the 587 evergreen Q. calliprinos. The Q. boissieri ARQts values during winter exfoliation were 588 significantly higher than the Q. calliprinos (0.51 ± 0.02 vs. 0.30 ± 0.02, P < 0.0001, Welch test), 589 while during the foliated period higher Q. boissieri values were not significant (0.41 ± 0.03 vs. 590 0.34 ± 0.03, P = 0.11, t test). In contrast, the species did not differ in their ARQts24 values (P = 591 0.66, t test) also during winter exfoliation (P = 0.07, Welch test). 592 The seasonal variability of the Q. boissieri was explained by physical factors in the 593 backward selection technique. The most effective factor was the accumulated rain in the 3 weeks 2 2 594 prior to sampling for ARQts (R = 0.85, P < 0.001) and for ARQts24 (R = 0.93, P < 0.001). Soil 2 2 595 moisture was also correlated with ARQts (R = 0.71, P = 0.04) and with ARQts24 (R = 0.82, 596 P = 0.03). One outlier point was excluded in each correlation. For the Q. calliprinos no

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

597 correlation was found. Inverse correlation with marginal significance was found between the Q. 2 598 boissieri ARQts at 20 cm above the ground and ARQsa (R = 0.41, P = 0.06) after excluding 1 599 outlier point out of 10 (measured in April 2017 when ARQts was minimal, Fig. 3).

600

601 Figure 7. Scatter plot of stem ARQ (ratio of CO2 efflux/O2 influx) measured from incubated 602 stem cores containing phloem and cambium tissues (ARQts) sampled 20 (filled symbols) and 130 603 cm (hollow symbols) above ground from the main trunks of the tree species Quercus calliprinos 604 and Quercus boissieri, against soil air ARQ measured below the same trees. Each point 605 represents the mean value of three trees (stems and underlying soils) measured in each campaign. 606 The P values of the correlations are 0.7393, 0.0618, 0.8973, and 0.533 ordered as appears in the 607 legend.

608 4 Discussion 609 In the first part of the discussion we interpret the results of the current study. The second part is 610 dedicated to discussion about methodology and future research towards more meaningful use of 611 ARQ in soil and ecosystem studies.

612 4.1 Bulk-soil ARQ results indicate the respiratory substrate is more reduced than 613 mean SOC

614 The overall weighted mean of the dissolution-corrected ARQbs is 0.75, well within the range of 615 previous ARQbs assessments (Fig. 1) and similar to values of 0.75-0.80 measured in different 616 natural soils in Germany [Dilly, 2001; 2003; Dilly and Zyakun, 2008]. An ARQbs value of 0.75 is 617 appreciably below 0.95, the ARQbs expected by SOM-Cox median value in soils meta-analysis 618 study, and slightly below 0.77, the lowest expected value [Worrall et al., 2013]. Several potential 619 processes can be ruled out as drivers for the low weighted mean ARQbs, including dissolution of 620 respired CO2 in the soil water (after the dissolution correction) and reversible effects such O2 621 uptake by oxidizing Fe2+ or Mn2+ that should be canceled by anaerobic respiration that must 622 occur at different time in the soil (e.g. to reduce Fe and Mn). However, other processes with 623 potentially irreversible effects, including microbial re-fixation of respired CO2 and oxidative 624 depolymerization are both possible explanations.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

625 Studies of CO2 ‘dark’ fixation showed that rates scaled with microbial biomass regardless 626 of soil depth [Akinyede et al., 2020]; thus we speculate that a constant fraction of respired CO2 is 627 re-fixed regardless the season. If the re-fixation products, which are usually relatively oxidized, 628 are mineralized shortly after synthesis, no net effect on ARQ is to be expected. However, if the 629 re-fixed C persists in microbial biomass or SOM, there will be a net decrease in ARQ. 630 Nonetheless, the effect is small and accounting for the highest observed re-fixation rates (5.6%; 631 [Akinyede et al., 2020]), only slightly increases our ARQbs mean, to 0.79. 632 Oxidative depolymerization breaks down macromolecules by introduction of oxygen-rich 633 groups, which forms small molecules that microbes can uptake easily. The polymer oxidation 634 step consumes O2 and is expected to lower ARQbs. However, the mineralization of the highly 635 oxygenated produced molecules is expected to have high ARQbs, with overall a zero net effect on 636 ARQbs. Net reduction of ARQbs might occur if the oxygen-rich molecules escape microbial 637 digestion and associate with minerals, where an extended time period may pass until their 638 mineralization [Kleber et al., 2015]. We estimate that in natural soils the fraction of C 639 associating to mineral surface is very small comparing to overall mineralization rates, thus the 640 net effect of oxidative depolymerization on ARQbs should be small. We therefore conclude the 641 mean ARQbs of 0.75 mainly reflects the stoichiometry of the respiratory substrate.

642 The difference between the expected ARQbs value according to the apparent SOM-Cox 643 and the measured ARQbs suggests that the overall SOM is more oxidized than the smaller SOM 644 pool that is more accessible to microbes and directly supports respiration. Since oxidized 645 compounds are energetically more labile than reduced compounds (Table 1) the plethora of low 646 ARQbs results indicate that Cox alone is insufficient to predict the decomposability of soil 647 compounds. The Cox difference between the respiratory substrate and the total SOM is similar to 648 radiocarbon measurements where bulk-soil respired CO2 is much younger than the total SOM 649 [Trumbore, 2000], which suggests most SOM has slow turnover rates with small contribution to 650 respiration. Hence, rather than representing the total SOM stoichiometry, ARQbs provides 651 information about the Cox of the actual labile pool utilized by microbes.

652 ARQbs values lower than expected based on the bulk-SOM composition also indicates an 653 oxidation of the remaining, unrespired SOM (unless the C influx to the SOM has exactly same 654 Cox as the respired C). This consequence is expected regardless the reason for the low ARQbs. If 655 reduced compounds with smaller O content are decomposed preferentially, the remainder SOM 656 is becoming more enriched with oxidized molecules. Same net effect is expected if the low 657 ARQbs is due to excess of bulk-soil O2 uptake. This expectation is in line with observed SOM 658 oxidation trends over longer time scales. For example, Rock-Eval indices show increases in Cox 659 (higher oxygen index and lower index) with soil depth [Sebag et al., 2016], with aging 660 of bare fallow [Barré et al., 2016], and with experimental soil warming [Poeplau et al., 2019]. 661 Free-air CO2 enrichment experiment (FACE) experiments also resulted in increased Cox values in 662 SOM after nine years of elevated CO2 treatment [Hockaday et al., 2015]. The oxidation was 663 explained by a loss of reduced and lignin and increased abundance of oxidized carbonyl 664 groups. Using the change in the C stocks, the authors further calculated the net respiration fluxes 665 accompanying the Cox change (OR = 1.477), would be equivalent to an ARQ of 0.68, which is 666 similar to our mean ARQbs value of 0.75.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

2+ 667 4.2 Fe redox changes affect ARQbs only at low respiration rates

668 ARQbs values were associated with respiration (i.e. O2 uptake) rate, which varied by two orders -1 669 of magnitude in our experiments (Fig. 4, 5a). At respiration rates above ~3 nanomole O2 g.DW -1 670 min most ARQbs values were similar to the mean weighted seasonal value (0.7-0.8) with 671 relatively small variability, while during lower respiration rates larger ARQbs variability were 672 observed (0.4-0.9; Fig. 4). One possible explanation for this observation is greater sensitivity of 673 the O2 measurement for smaller O2 fluxes. However, the low ARQbs values measured at low 674 temperatures (Fig. 3) and in the soil drying experiment (Fig. 5a) seem like part of a trend and not 675 just arbitrary analytical noise. Alternatively, the ARQbs and O2 uptake relation can be explained 676 by a shift in the dominant processes. When respiration rates are low, slower processes (i.e. those 677 normally contributing only in a minor way to larger O2 fluxes), such as oxidative 678 depolymerization and redox related changes in Fe, can have a measurable effect on ARQbs. As 679 respiration rates increase the ARQbs value related to substrate-stoichiometry will drown out other 680 effects.

681 In the jar incubations with soils recovering from anaerobic conditions ([O2] ~1%) the 2+ 682 ARQbs was 0.37 and [Fe ] decreased sharply. The stoichiometry for the overall oxidation of 2+ 2+ + 683 Fe ions by O2, O2(aq) + 4Fe + 6H2O ↔ 4FeOOH(s) + 8H [Burke and Banwart, 2002], 684 explains 27% of the drop in [O2], another third of the O2 uptake can be explained by faster 685 oxidation of soil organic matter that usually follows anaerobic conditions (e.g. [Keiluweit et al., 686 2017]), while the last third can be explained by the same microbial respiration as in the control 687 soils. For comparison, the initial [O2] in the control jars was ~7% and the subsequent ARQbs was 2+ 688 0.74 with no change in [Fe ]. As the air in the anaerobic treatment was nearly anoxic, the ARQbs 689 difference between 0.74 and 0.37 seems to represent the maximal effect of Fe2+ oxidation for the 690 site. However, the soil drying-rewetting experiment indicates Fe redox change has only a small th 691 influence under field conditions (Fig. 5). The decrease in ARQbs values on the 11 day after soil 692 wetting seems to be the result of Fe2+ oxidation that occurred around the same time – by this time 693 point, respiration rates were already low. We estimated that the amount of O2 decrease due to 694 Fe2+ oxidation, which is equivalent to the amount of alternative oxidants during anaerobic 695 respiration, is less than 10% of the O2 flux when respiration rates were higher. In addition, the 696 minimal [O2] measured in the soil air was 18.14% therefore anoxia might present only in 697 microsites and not in the whole soil profile. Moreover, the fast [O2] transition from ~1% to 698 20.95% in the jar incubations that resulted the sharp ARQbs decrease is not expected in the field. 699 Gradual [O2] increase with smaller effect on ARQbs is rather expected. Thus, we conclude that 700 redox-related O2 consumption is significant at Odem forest only at low respiration rates.

701 4.3 Seasonality of soil air ARQ indicates soil respiration is controlled by bulk-soil 702 respiration

703 The overall mean ARQsa was 0.76, within the range of previous studies (Fig. 1). ARQsa was 704 almost always higher than ARQbs and when ARQroot was measured its values exceeded ARQsa, 705 apparently confirming that ARQsa value is the weighted mean of those two end members and that 706 ARQ can be used for partitioning of soil respiration sources (Fig. 6). Our results further support 707 the view that variability in ARQsa reflects shifting dominance between root and microbially- 708 mediated heterotrophic respiration sources [Hicks Pries et al., 2019]. 709 The backward selection technique indicates that on a yearly basis water-related 710 parameters are the main factors controlling the seasonal ARQsa variability in this Mediterranean

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

711 site, with negligible effect of Tsoil (Fig. 2a; Eq. 7). With two orders of magnitude variability, the 712 bulk-soil O2 uptake rate (at room temperature) was also mainly controlled by soil moisture (Fig 713 2d; Eq. 6). We lack information about the seasonal variability of root respiration rates, but based 714 on other studies in Mediterranean climates, it can be assumed the rate from tree roots is rather 715 constant [Tang and Baldocchi, 2005; Vose and Ryan, 2002]. We therefore estimate that in Odem 716 forest ARQsa is mainly controlled by variability in ARQbs, which in turn mainly controlled by 717 soil moisture. Tsoil was important factor with positive effect on ARQsa only outside the high 718 growth period (May to January) when soil respiration rates were slow (Fig. 2). A similar positive o 719 effect of Tsoil on ARQsa was observed at 30 cm depth in heated soils (+4 C) during winter [Hicks 720 Pries et al., 2019]. This trend is in accordance with the effect of temperature on ARQbs; a 721 positive effect when respiration rates are slow, and no effect when respiration rates are high (Fig. 722 3, 4).

723 Deviations of ARQsa from ARQbs and ARQroot values might resulted from soil profile 724 processes that might be present in the field and not in the ARQbs and ARQroot incubations. For 725 example, CO2 that otherwise would be released to the soil air might dissolve in the soil water and 726 leach. This process can explain the few ARQsa values that were equal to or lower than ARQbs 727 during the second and wetter winter (Fig. 2, 6). In contrast if the soil water does not leach the 728 dissolved CO2 can degasses back to the pore space after water evaporation or warming that 729 reduces CO2 solubility. Such degassing could explain the spike in ARQsa observed during the 730 last campaign of the second winter (Fig 6). In the area of Odem forest only 10-30% of annual 731 precipitation (950 mm) leaches to groundwater [Dafny et al., 2006], most of it during episodic 732 intensive rain events during winter. Therefore we estimate the loss of respired CO2 to 733 groundwater is negligible on an annual basis, though it might affect specific sampling dates.

734 4.4 Tree stems ARQ 735 The ARQ values measured in tree stem tissues were considerably lower than 1.0, the value 736 expected from carbohydrate respiration (Table 1) in most plants [Hoch et al., 2003; Plaxton and 737 Podestá, 2007]. The mean ARQts and ARQts24 values were 0.39 and 0.68, respectively. Damage 738 during the tissue extraction from the stems might result in a burst of O2 uptake that lowered 739 ARQts values. One indication for such a short-term wound response on O2 uptake is elevated 740 H2O2 production observed within two hours after epicormic shoot wounding [Tian et al., 2015]. 741 Therefore, ARQts24 (measured 24 h after tissues extraction) can be considered as less prone to 742 artifacts, although metabolic change during the 24 h period is possible [Hilman et al., 2019]. The 743 ARQts24 mean values of ~0.7 could be explained by pure lipid substrates for respiration (Table 744 1). This is, however, not expected, especially in the tree genera Quercus [Hoch et al., 2003; 745 Sinnott, 1918]. The current results thus reinforce our recent studies and suggest that tree stem 746 ARQ do not merely reflect the substrate’s stoichiometry and additional non-respiratory effects 747 are at work [Angert et al., 2012; Hilman et al., 2019]. 748 Our observations indicate differences between the species that are related to phenology. 749 During the foliated period of the deciduous Q. boissieri its ARQts was similar to the evergreen Q. 750 calliprinos values, however during winter exfoliation the Q. boissieri values increased in ~0.2 751 ARQ units above the Q. calliprinos values (Fig. 2). The seasonal change of ARQts and ARQts24 752 of the Q. boissieri were explained by water related factors suggesting low tree stem ARQ is 753 associated with dry conditions. We speculated that if re-fixation of respired CO2 by the enzyme 754 PEPC is the process that lowers tree-stems ARQ, the enzyme’s products like the organic acids

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

755 citrate and malate might be exported via the phloem to the roots and be secreted to the soil as 756 root exudates [Hoffland et al., 2006; Shane et al., 2004]. In agreement, a recent study in a 757 different Mediterranean forest in Israel reported that root exudation rates were highest in the dry 758 season and were associated with high soil temperature and low soil moisture [Jakoby et al., 759 2020]. Indeed, the inverse relationship between ARQts vs. ARQsa at 20 cm in the deciduous Q. 760 boissieri (Fig. 7) might suggest catabolism of organic acids, with high ARQ in the soil offsetting 761 low stem ARQ (Table 1). However the Q. calliprinos ARQts values were rather uniform 762 throughout the year without correlation with ARQsa. The fact that ARQsa of both species had 763 almost the same seasonal changes suggests that the variability in ARQsa is not related to ARQts. 764 It is highly probable that our tube sampling from depth of ~15 cm, used to estimate ARQsa, did 765 not fully capture the autotrophic signal of shallow fine roots. Section 4.5.2 adds more discussion 766 about this topic.

767 4.5 What should be done next? 768 In this section we discuss future research of ARQ in soils. Suggested future steps in tree stem 769 ARQ research are discussed at [Hilman et al., 2019].

770 4.5.1 Bulk-soil incubations

771 For estimating ARQbs we measured the change of [CO2] and [O2] in the headspace of closed- 772 system incubations. We further corrected for CO2 dissolution in the soil moisture, especially due 773 to the relatively basic pH value (6.6) and large amount of soil water per headspace volume. A 774 way to reduce the amount of dissolved CO2 is by increasing the incubation set-up. With fixed 775 soil water content, increasing the ratio incubation headspace/ soil mass by a given factor would 776 decrease the amount of dissolved CO2 by the same factor.

777 We conclude that when respiration rates are high, the ARQbs signal mostly reflects the 778 substrate stoichiometry of microbial respiration. One way to test this would be to amend soil 779 incubations with compounds differing in their expected ARQ, for example lipids or soluble 780 amino acids with ARQ < 1, sugars with ARQ = 1, and organic-acids with ARQ >1. Ideally the 781 compounds should have different δ13C signatures than the SOM to allow calculation of the 782 contribution of the amended compound to the respired C. Previously, addition of sugars (with 783 ARQ = 1) to bulk-soil incubations indeed resulted in higher ARQbs compared to basal values of 784 0.8, but the resulting ARQ values varied between 0.95 and 1.6 and depended on the added 785 amount of sugar and the sampling timing after the amendment [Dilly and Zyakun, 2008]. High 786 amounts of added substrate caused a ‘priming effect’ that rapidly altered microbial composition 787 and metabolism, thus the resulting ARQ does not represent purely the stoichiometry of the 788 amendment. We therefore suggest amending soils with smaller amounts of 13C-labeled substrate 789 to better mimic natural input rates. The role of oxidative depolymerization and CO2 re-fixation in 790 ARQbs variability can be tested by parallel measurement of ARQbs and the activity of oxidases 791 (enzymes that use O to break-down large molecules) and rates of CO2 re-fixation.

792 Temperature was found to have an effect on ARQbs. Moisture effects were not tested o 793 separately, but the higher ARQbs at 22 and 30 C in Q. calliprinos compared to Q. boissieri soils 794 may reflect the higher soil moisture in the Q. calliprinos (Fig. 3). Higher water content can affect 795 ARQbs via faster respiration (similarly to the discussed temperature effect), or via faster diffusion 796 of dissolved organic C that can fuel microbial respiration with different substrates. Better 797 understanding of these effects can also improve ARQbs predictions with seasonal changes.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

798 Therefore factorial incubations in different temperatures and soil moistures to test these effects 799 separately are recommended.

800 4.5.2 Total soil ARQ and portioning respiration sources

801 As mentioned in the introduction, the very few available soil-chamber-ARQsa measurements 802 yielded higher values than estimated from gas profiles and diffusion modeling. The assumed 803 reason is that the soil chamber better captures the signal of processes like respiration from fine 804 roots (with high ARQ) that are usually concentrated in soil horizons above the ones sampled by 805 gas probes (e.g. [Claus and George, 2005]). Under-representation of the shallow roots might 806 explain the lack of the expected inverse stem-soil respiration ARQ relations. The influence of 807 surface litter decomposition is another process not represented in gas diffusion models. Hence, to 808 evaluate the contribution of soil respiration ARQ to the total ecosystems ARQ balance requires 809 surface chamber measurements to at least compare with gas tube/diffusion estimates. However, 810 measurement of ARQ using tube sampling is easier than chamber sampling since the O2 signals 811 are larger and easier to detect, and the sampling is fast and simple technically. Comparing both 812 measurements is expected to yield interesting results, especially testing the assumptions that are 813 presented in Section 2.3 and Eq. 1, and to make a sensitivity test to find under which conditions 814 (e.g. advective mixing) Eq. 1 becomes invalid. 815 ARQ has the potential to partition the sources of soil respiration. In comparison to other 816 non-destructive methods like the bomb-radiocarbon approach and canopy δ13C-labeling the use 817 of ARQ can be technically simpler and more cost-effective [Kuzyakov, 2006]. There are few 818 considerations that yet should be taken: different ARQbs signals from soils sampled from 819 different depths, their temporal variability, and their relative contributions to the total respiration 820 fluxes.

821 Acknowledgments 822 We thank Goni Shachar for helping with field work, Maayan Farbman for helping in soil 823 incubations, Joe Von Fischer for valuable discussion, and Susan Trumbore for critical reading. 824 This research was supported by the Israel Science Foundation (grant # 773/19), and by the 825 German–Israeli Foundation for Scientific Research and Development (no. 1334/2016). The data 826 are available in https://figshare.com/s/db7b0e2c5b5dfbcb28bf.

827 References

828 Akinyede, R., M. Taubert, M. Schrumpf, S. Trumbore, and K. Küsel (2020), Rates of dark CO2 829 fixation are driven by microbial biomass in a temperate forest soil, Soil Biology and 830 Biochemistry, 150, 107950. 831 Angert, A., J. Muhr, R. Negron Juarez, W. Alegria Muñoz, G. Kraemer, J. Ramirez Santillan, E. 832 Barkan, S. Mazeh, J. Q. Chambers, and S. E. Trumbore (2012), Internal respiration of Amazon 833 tree stems greatly exceeds external CO2 efflux, Biogeosciences, 9(12), 4979-4991, 834 doi:10.5194/bg-9-4979-2012. 835 Angert, A., and Y. Sherer (2011), Determining the relationship between tree-stem respiration and 836 CO2 efflux by δO2/Ar measurements, Rapid Commun Mass Spectrom, 25(12), 1752-1756, 837 doi:10.1002/rcm.5042.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

838 Angert, A., D. Yakir, M. Rodeghiero, Y. Preisler, E. Davidson, and T. Weiner (2015), Using O2 839 to study the relationships between soil CO2 efflux and soil respiration, Biogeosciences, 12(7), 840 2089-2099. 841 Aon, M. A., D. E. Sarena, J. L. Burgos, and S. Cortassa (2001a), Interaction between gas 842 exchange rates, physical and microbiological properties in soils recently subjected to agriculture, 843 Soil and Tillage Research, 60(3-4), 163-171, doi:10.1016/s0167-1987(01)00191-x. 844 Aon, M. A., D. E. Sarena, J. L. Burgos, and S. Cortassa (2001b), (Micro)biological, chemical 845 and physical properties of soils subjected to conventional or no-till management: an assessment 846 of their quality status, Soil and Tillage Research, 60(3-4), 173-186, doi:10.1016/s0167- 847 1987(01)00190-8.

848 Aubrey, D. P., and R. O. Teskey (2009), Root-derived CO2 efflux via xylem stream rivals soil 849 CO2 efflux, New Phytol, 184(1), 35-40, doi:10.1111/j.1469-8137.2009.02971.x. 850 Bais, H. P., T. L. Weir, L. G. Perry, S. Gilroy, and J. M. Vivanco (2006), The role of root 851 exudates in rhizosphere interations with plants and other organisms, in Annual Review of Plant 852 Biology, edited, pp. 233-266, Annual Reviews, Palo Alto, 853 doi:doi:10.1146/annurev.arplant.57.032905.105159. 854 Barré, P., et al. (2016), The energetic and chemical signatures of persistent soil organic matter, 855 Biogeochemistry, 130(1-2), 1-12, doi:10.1007/s10533-016-0246-0.

856 Battle, M. O., et al. (2019), Atmospheric measurements of the terrestrial O2 : CO2 exchange ratio 857 of a midlatitude forest, Atmospheric Chemistry and Physics, 19(13), 8687-8701, 858 doi:10.5194/acp-19-8687-2019. 859 Benavente, J., I. Vadillo, F. Carrasco, A. Soler, C. Liñán, and F. Moral (2010), Air Carbon 860 Dioxide Contents in the Vadose Zone of a Mediterranean Karst, Vadose Zone Journal, 9(1), 126- 861 136, doi:10.2136/vzj2009.0027. 862 Berveiller, D., and C. Damesin (2007), Carbon assimilation by tree stems: potential involvement 863 of phosphoenolpyruvate carboxylase, Trees, 22(2), 149-157, doi:10.1007/s00468-007-0193-4. 864 Berveiller, D., J. Vidal, J. Degrouard, F. Ambard-Bretteville, J. N. Pierre, D. Jaillard, and C. 865 Damesin (2007), Tree stem phosphoenolpyruvate carboxylase (PEPC): lack of biochemical and 866 localization evidence for a C4-like photosynthesis system, New Phytol, 176(4), 775-781, 867 doi:10.1111/j.1469-8137.2007.02283.x. 868 Bloemen, J., M. A. McGuire, D. P. Aubrey, R. O. Teskey, and K. Steppe (2013), Transport of 869 root-respired CO(2) via the transpiration stream affects aboveground carbon assimilation and 870 CO(2) efflux in trees, New Phytol, 197(2), 555-565, doi:10.1111/j.1469-8137.2012.04366.x. 871 Bloom, A. J., R. M. Caldwell, J. Finazzo, R. L. Warner, and J. Weissbart (1989), Oxygen and 872 carbon dioxide fluxes from barley shoots depend on nitrate assimilation, Plant Physiol, 91(1), 873 352-356, doi:10.1104/pp.91.1.352. 874 Bowling, D. R., J. E. Egan, S. J. Hall, and D. A. Risk (2015), Environmental forcing does not 875 induce diel or synoptic variation in the carbon isotope content of forest soil respiration, 876 Biogeosciences, 12(16), 5143-5160, doi:10.5194/bg-12-5143-2015. 877 Bowman, W. P., M. M. Barbour, M. H. Turnbull, D. T. Tissue, D. Whitehead, and K. L. Griffin 878 (2005), Sap flow rates and sapwood density are critical factors in within- and between-tree

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

879 variation in CO2 efflux from stems of mature Dacrydium cupressinum trees, New Phytol, 167(3), 880 815-828, doi:10.1111/j.1469-8137.2005.01478.x. 881 Burke, S. P., and S. A. Banwart (2002), A geochemical model for removal of iron(II)(aq) from 882 mine water discharges, Applied Geochemistry, 17(4), 431-443, doi:10.1016/s0883- 883 2927(01)00092-0. 884 Chen, D., L. Zhou, X. Rao, Y. Lin, and S. Fu (2010), Effects of root diameter and root nitrogen 885 concentration on in situ root respiration among different seasons and tree species, Ecological 886 Research, 25(5), 983-993, doi:10.1007/s11284-010-0722-2. 887 Claus, A., and E. George (2005), Effect of stand age on fine-root biomass and biomass 888 distribution in three European forest chronosequences, Canadian Journal of Forest Research, 889 35(7), 1617-1625, doi:10.1139/x05-079. 890 Cuezva, S., A. Fernandez-Cortes, D. Benavente, P. Serrano-Ortiz, A. S. Kowalski, and S. 891 Sanchez-Moral (2011), Short-term CO2(g) exchange between a shallow karstic cavity and the 892 external atmosphere during summer: Role of the surface soil layer, Atmospheric Environment, 893 45(7), 1418-1427, doi:10.1016/j.atmosenv.2010.12.023. 894 Dafny, E., A. Burg, and H. Gvirtzman (2006), Deduction of groundwater flow regime in a 895 basaltic aquifer using geochemical and isotopic data: The Golan Heights, Israel case study, 896 Journal of Hydrology, 330(3-4), 506-524, doi:10.1016/j.jhydrol.2006.04.002. 897 Davidson, G. R. (1995), The Stable Isotopic Composition and Measurement of Carbon in Soil 898 CO2, Geochimica Et Cosmochimica Acta, 59(12), 2485-2489, doi:Doi 10.1016/0016- 899 7037(95)00143-3. 900 De Vries, F. W. T. P., A. H. M. Brunsting, and H. H. Van Laar (1974), Products, requirements 901 and efficiency of biosynthesis a quantitative approach, Journal of Theoretical Biology, 45(2), 902 339-377, doi:10.1016/0022-5193(74)90119-2. 903 Desrochers, A., S. M. Landhausser, and V. J. Lieffers (2002), Coarse and fine root respiration in 904 aspen (Populus tremuloides), Tree Physiol, 22(10), 725-732, doi:10.1093/treephys/22.10.725. 905 Dilly, O. (2001), Microbial respiratory quotient during basal metabolism and after glucose 906 amendment in soils and litter, Soil Biology and Biochemistry, 33(1), 117-127, 907 doi:10.1016/s0038-0717(00)00123-1. 908 Dilly, O. (2003), Regulation of the respiratory quotient of soil microbiota by availability of 909 nutrients, FEMS Microbiol Ecol, 43(3), 375-381, doi:10.1111/j.1574-6941.2003.tb01078.x. 910 Dilly, O., and A. Zyakun (2008), Priming Effect and Respiratory Quotient in a Forest Soil 911 Amended with Glucose, Geomicrobiology Journal, 25(7-8), 425-431, 912 doi:10.1080/01490450802403099. 913 Druschel, G. K., D. Emerson, R. Sutka, P. Suchecki, and G. W. Luther (2008), Low-oxygen and 914 chemical kinetic constraints on the geochemical niche of neutrophilic iron(II) oxidizing 915 microorganisms, Geochimica et Cosmochimica Acta, 72(14), 3358-3370, 916 doi:10.1016/j.gca.2008.04.035. 917 Egan, J. E., D. R. Bowling, and D. A. Risk (2019), Technical Note: Isotopic corrections for the 918 radiocarbon composition of CO2 in the soil gas environment must account for diffusion and 919 diffusive mixing, Biogeosciences, 16(16), 3197-3205, doi:10.5194/bg-16-3197-2019.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

920 Emmerich, W. E. (2003), Carbon dioxide fluxes in a semiarid environment with high carbonate 921 soils, Agricultural and Forest Meteorology, 116(1-2), 91-102, doi:10.1016/s0168- 922 1923(02)00231-9. 923 Gallagher, M. E., F. L. Liljestrand, W. C. Hockaday, and C. A. Masiello (2017), Plant species, 924 not climate, controls aboveground biomass O2:CO2exchange ratios in deciduous and coniferous 925 ecosystems, J Geophys Res-Biogeo, 122(9), 2314-2324, doi:10.1002/2017jg003847. 926 Gallagher, T. M., and D. O. Breecker (2020), The Obscuring Effects of Calcite Dissolution and 927 Formation on Quantifying Soil Respiration, Global Biogeochemical Cycles, 34(12), 928 e2020GB006584, doi:ARTN e2020GB006584 929 10.1029/2020GB006584. 930 Gross, A., and A. Angert (2017), Use of 13C‐and phosphate 18O‐labeled substrate for studying 931 phosphorus and carbon cycling in soils: a proof of concept, Rapid Commun Mass Sp, 31(11), 932 969-977. 933 Grossiord, C., L. Mareschal, and D. Epron (2012), Transpiration alters the contribution of 934 autotrophic and heterotrophic components of soil CO2 efflux, New Phytol, 194(3), 647-653, 935 doi:10.1111/j.1469-8137.2012.04102.x. 936 Hall, S. J., and W. L. Silver (2013), Iron oxidation stimulates organic matter decomposition in 937 humid tropical forest soils, Glob Chang Biol, 19(9), 2804-2813, doi:10.1111/gcb.12229. 938 Harned, H. S., and R. Davis (1943), The ionization constant of carbonic acid in water and the 939 solubility of carbon dioxide in water and aqueous salt from 0 to 50 degrees, Journal of 940 the American Chemical Society, 65(10), 2030-2037, doi:DOI 10.1021/ja01250a059. 941 Harned, H. S., and S. R. Scholes Jr (1941), The ionization constant of HCO3-from 0 to 50, 942 Journal of the American Chemical Society, 63(6), 1706-1709. 943 Hawkins, H.-J., M. D. Cramer, and E. George (1999), Root respiratory quotient and nitrate 944 uptake in hydroponically grown non-mycorrhizal and mycorrhizal wheat, Mycorrhiza, 9(1), 57- 945 60. 946 Hibberd, J. M., and W. P. Quick (2002), Characteristics of C4 photosynthesis in stems and 947 petioles of C3 flowering plants, Nature, 415(6870), 451-454, doi:10.1038/415451a. 948 Hicks Pries, C., A. Angert, C. Castanha, B. Hilman, and M. S. Torn (2019), Using Respiration 949 Quotients to Track Changing Sources of Soil Respiration Seasonally and with Experimental 950 Warming, Biogeosciences Discuss., 2019, 1-20, doi:10.5194/bg-2019-232.

951 Hilman, B., et al. (2019), Comparison of CO2 and O2 fluxes demonstrate retention of respired 952 CO2 in tree stems from a range of tree species, Biogeosciences, 16(1), 177-191, doi:10.5194/bg- 953 16-177-2019. 954 Hoch, G., A. Richter, and C. Korner (2003), Non-structural carbon compounds in temperate 955 forest trees, Plant, Cell and Environment, 26(7), 1067-1081, doi:10.1046/j.0016- 956 8025.2003.01032.x. 957 Hockaday, W. C., M. E. Gallagher, C. A. Masiello, J. A. Baldock, C. M. Iversen, and R. J. Norby 958 (2015), Forest soil carbon oxidation state and oxidative ratio responses to elevated CO2, Journal 959 of Geophysical Research-Biogeosciences, 120(9), 1797-1811, doi:10.1002/2015jg003010.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

960 Hoffland, E., R. Boogaard, J. Nelemans, and G. Findenegg (2006), Biosynthesis and root 961 exudation of citric and malic acids in phosphate-starved rape plants, New Phytologist, 122(4), 962 675-680, doi:10.1111/j.1469-8137.1992.tb00096.x. 963 Hoffland, E., R. Van Den Boogaard, J. Nelemans, and G. Findenegg (1992), Biosynthesis and 964 root exudation of citric and malic acids in phosphate-starved rape plants, New Phytologist, 965 122(4), 675-680, doi:https://doi.org/10.1111/j.1469-8137.1992.tb00096.x. 966 Ishidoya, S., S. Murayama, C. Takamura, H. Kondo, N. Saigusa, D. Goto, S. Morimoto, N. Aoki, 967 S. Aoki, and T. Nakazawa (2013), O2: CO2 exchange ratios observed in a cool temperate 968 deciduous forest ecosystem of central Japan, Tellus B: Chemical and Physical Meteorology, 969 65(1), 21120. 970 Jakoby, G., I. Rog, S. Megidish, and T. Klein (2020), Enhanced root exudation of mature 971 broadleaf and conifer trees in a Mediterranean forest during the dry season, Tree Physiol, 40(11), 972 1595-1605, doi:10.1093/treephys/tpaa092. 973 Kaplan, D., and M. Gutman (1996), Effect of Thinning and Grazing on Tree Development and 974 the Visual Aspect of an Oak Forest on the Golan Heights, Israel Journal of Plant Sciences, 975 44(4), 381-386, doi:10.1080/07929978.1996.10676659. 976 Keeling, R. F. (1988), Development of an interferometric oxygen analyzer for precise 977 measurement of the atmospheric O2 mole fraction, Harvard University, Cambridge, 978 Massachusetts. 979 Keiluweit, M., K. Gee, A. Denney, and S. Fendorf (2018), Anoxic microsites in upland soils 980 dominantly controlled by clay content, Soil Biology and Biochemistry, 118, 42-50, 981 doi:10.1016/j.soilbio.2017.12.002. 982 Keiluweit, M., T. Wanzek, M. Kleber, P. Nico, and S. Fendorf (2017), Anaerobic microsites 983 have an unaccounted role in soil carbon stabilization, Nature communications, 8(1), 1771. 984 Kleber, M., K. Eusterhues, M. Keiluweit, C. Mikutta, R. Mikutta, and P. S. Nico (2015), 985 Mineral–organic associations: formation, properties, and relevance in soil environments, in 986 Advances in agronomy, edited, pp. 1-140, Elsevier. 987 Kuzyakov, Y. (2006), Sources of CO2 efflux from soil and review of partitioning methods, Soil 988 Biology and Biochemistry, 38(3), 425-448, doi:https://doi.org/10.1016/j.soilbio.2005.08.020. 989 Lambers, H., F. S. Chapin, and T. L. Pons (2008), Respiration, in Plant Physiological Ecology, 990 edited, pp. 101-150, Springer New York, New York, NY, doi:10.1007/978-0-387-78341-3_3. 991 LaRowe, D. E., and P. Van Cappellen (2011), Degradation of natural organic matter: a 992 thermodynamic analysis, Geochimica et Cosmochimica Acta, 75(8), 2030-2042. 993 Liptzin, D., and W. L. Silver (2009), Effects of carbon additions on iron reduction and 994 phosphorus availability in a humid tropical forest soil, Soil Biology and Biochemistry, 41(8), 995 1696-1702.

996 Ma, J., Z. Y. Wang, B. A. Stevenson, X. J. Zheng, and Y. Li (2013), An inorganic CO2 diffusion 997 and dissolution process explains negative CO2 fluxes in saline/alkaline soils, Sci Rep, 3, 2025, 998 doi:10.1038/srep02025.

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

999 Masiello, C. A., M. E. Gallagher, J. T. Randerson, R. M. Deco, and O. A. Chadwick (2008), 1000 Evaluating two experimental approaches for measuring ecosystem carbon oxidation state and 1001 oxidative ratio, Journal of Geophysical Research, 113(G3), doi:10.1029/2007jg000534.

1002 Massman, W. J. (1998), A review of the molecular diffusivities of H2O, CO2, CH4, CO, O3, SO2, 1003 NH3, N2O, NO, and NO2 in air, O2 and N2 near STP, Atmospheric Environment, 32(6), 1111- 1004 1127, doi:10.1016/s1352-2310(97)00391-9. 1005 Miltner, A., F. D. Kopinke, R. Kindler, D. E. Selesi, A. Hartmann, and M. Kastner (2005), Non- 1006 phototrophic CO2 fixation by soil microorganisms, Plant and Soil, 269(1-2), 193-203, 1007 doi:10.1007/s11104-004-0483-1. 1008 Plaxton, W. C., and F. E. Podestá (2007), The Functional Organization and Control of Plant 1009 Respiration, Critical Reviews in Plant Sciences, 25(2), 159-198, 1010 doi:10.1080/07352680600563876. 1011 Poeplau, C., P. Barré, L. Cécillon, F. Baudin, and B. D. Sigurdsson (2019), Changes in the Rock- 1012 Eval signature of soil organic carbon upon extreme soil warming and chemical oxidation-A 1013 comparison, Geoderma, 337, 181-190. 1014 Pregitzer, K. S., M. J. Laskowski, A. J. Burton, V. C. Lessard, and D. R. Zak (1998), Variation in 1015 sugar maple root respiration with root diameter and soil depth, Tree Physiol, 18(10), 665-670, 1016 doi:10.1093/treephys/18.10.665. 1017 R Core Team (2019), R: A Language and Environment for Statistical Computing, edited. 1018 Rachmilevitch, S., H. Lambers, and B. Huang (2006), Root respiratory characteristics associated 1019 with plant adaptation to high soil temperature for geothermal and turf-type Agrostis species, J 1020 Exp Bot, 57(3), 623-631, doi:10.1093/jxb/erj047. 1021 Sanchez-Canete, E. P., G. A. Barron-Gafford, and J. Chorover (2018), A considerable fraction of 1022 soil-respired CO2 is not emitted directly to the atmosphere, Sci Rep, 8(1), 13518, 1023 doi:10.1038/s41598-018-29803-x. 1024 Schill, V., W. Hartung, B. Orthen, and M. H. Weisenseel (1996), The xylem sap of maple (Acer 1025 platanoides) trees—sap obtained by a novel method shows changes with season and height, 1026 Journal of Experimental Botany, 47(1), 123-133, doi:10.1093/jxb/47.1.123. 1027 Sebag, D., et al. (2016), Dynamics of soil organic matter based on new Rock-Eval indices, 1028 Geoderma, 284, 185-203, doi:10.1016/j.geoderma.2016.08.025. 1029 Seibt, U., W. A. Brand, M. Heimann, J. Lloyd, J. P. Severinghaus, and L. Wingate (2004), 1030 Observations of O2:CO2 exchange ratios during ecosystem , Global 1031 Biogeochemical Cycles, 18(4), n/a-n/a, doi:10.1029/2004gb002242.

1032 Severinghaus, J. P. (1995), Studies of the terrestrial O2 and carbon cycles in sand dune gases and 1033 in biosphere 2, Columbia University. 1034 Sexstone, A. J., N. P. Revsbech, T. B. Parkin, and J. M. Tiedje (1985), Direct measurement of 1035 oxygen profiles and denitrification rates in soil aggregates, Soil science society of America 1036 journal, 49(3), 645-651. 1037 Shane, M. W., M. D. Cramer, S. Funayama-Noguchi, G. R. Cawthray, A. H. Millar, D. A. Day, 1038 and H. Lambers (2004), Developmental physiology of cluster-root carboxylate synthesis and

ESSOAr | https://doi.org/10.1002/essoar.10506181.1 | CC_BY_4.0 | First posted online: Fri, 12 Feb 2021 11:04:44 | This content has not been peer reviewed.

Confidential manuscript submitted to JGR Biogeosciences

1039 exudation in harsh hakea. Expression of phosphoenolpyruvate carboxylase and the alternative 1040 oxidase, Plant Physiol, 135(1), 549-560, doi:10.1104/pp.103.035659. 1041 Sinnott, E. W. (1918), Factors determining character and distribution of food reserve in woody 1042 plants, Botanical Gazette, 66(2), 162-175. 1043 Stephens, B. B., P. S. Bakwin, P. P. Tans, R. M. Teclaw, and D. D. Baumann (2007), 1044 Application of a Differential Fuel-Cell Analyzer for Measuring Atmospheric Oxygen Variations, 1045 Journal of Atmospheric and Oceanic Technology, 24(1), 82-94, doi:10.1175/jtech1959.1. 1046 Stringer, J. W., and T. W. Kimmerer (1993), Refixation of xylem sap CO2 in Populus deltoides, 1047 Physiologia Plantarum, 89(2), 243-251. 1048 Stumm, W., and J. J. Morgan (1996), Aquatic chemistry : chemical equilibria and rates in 1049 natural waters, 3rd ed., xvi, 1022 p. pp., Wiley, New York. 1050 Tang, J., and D. D. Baldocchi (2005), Spatial–temporal variation in soil respiration in an oak– 1051 grass savanna ecosystem in California and its partitioning into autotrophic and heterotrophic 1052 components, Biogeochemistry, 73(1), 183-207, doi:10.1007/s10533-004-5889-6. 1053 Teskey, R. O., A. Saveyn, K. Steppe, and M. A. McGuire (2008), Origin, fate and significance of 1054 CO2 in tree stems, New Phytol, 177(1), 17-32, doi:10.1111/j.1469-8137.2007.02286.x. 1055 Tian, W. M., S. G. Yang, M. J. Shi, S. X. Zhang, and J. L. Wu (2015), Mechanical wounding- 1056 induced laticifer differentiation in rubber tree: An indicative role of dehydration, hydrogen 1057 peroxide, and jasmonates, J Plant Physiol, 182, 95-103, doi:10.1016/j.jplph.2015.04.010. 1058 Trumbore, S. (2000), Age of soil organic matter and soil respiration: Radiocarbon constraints on 1059 belowground C dynamics, Ecol Appl, 10(2), 399-411, doi:Doi 10.2307/2641102. 1060 Ubierna, N., A. S. Kumar, L. A. Cernusak, R. E. Pangle, P. J. Gag, and J. D. Marshall (2009), 13 1061 Storage and transpiration have negligible effects on δ C of stem CO2 efflux in large conifer 1062 trees, Tree Physiology, 29(12), 1563-1574. 1063 Vose, J. M., and M. G. Ryan (2002), Seasonal respiration of foliage, fine roots, and woody 1064 tissues in relation to growth, tissue N, and photosynthesis, Global Change Biology, 8(2), 182- 1065 193, doi:DOI 10.1046/j.1365-2486.2002.00464.x. 1066 Worrall, F., G. D. Clay, C. A. Masiello, and G. Mynheer (2013), Estimating the oxidative ratio 1067 of the global terrestrial biosphere carbon, Biogeochemistry, 115(1-3), 23-32, 1068 doi:10.1007/s10533-013-9877-6. 1069